chore: image_url to kb

alembic/versions/f6a1d6fc82d0_add_image_url_column_to_kb.py

@@ -0,0 +1,33 @@
+"""add: image url column to kb
+
+Revision ID: f6a1d6fc82d0
+Revises: 1cbb7317540a
+Create Date: 2025-12-01 16:55:06.113217
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+import sqlmodel.sql.sqltypes
+
+
+# revision identifiers, used by Alembic.
+revision: str = 'f6a1d6fc82d0'
+down_revision: Union[str, Sequence[str], None] = '1cbb7317540a'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('knowledge_chunk', sa.Column('image_url', sqlmodel.sql.sqltypes.AutoString(), nullable=True))
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column('knowledge_chunk', 'image_url')
+    # ### end Alembic commands ###

src/chatbot/models.py

@@ -31,5 +31,6 @@ class KnowledgeChunk(SQLModel, table=True):
     )
     chunk_index: int
     chunk_text: str
+    image_url: str | None = Field(default=None)
     embedding: List[float] = Field(sa_column=Column(Vector(768)))
     knowledge_base: "KnowledgeBase" = Relationship(back_populates="knowledge_chunk")

src/chatbot/router.py

@@ -57,7 +57,7 @@ async def semantic_search(
 
     sql = text(
         """
-        SELECT id, kb_id, chunk_text,
+        SELECT id, kb_id, chunk_text,image_url,
            embedding <#> :query_vec AS score
         FROM knowledge_chunk
         ORDER BY embedding <#> :query_vec ASC
@@ -75,6 +75,7 @@ async def semantic_search(
             chunk_id=str(r.id),
             kb_id=str(r.kb_id),
             text=r.chunk_text,
+            image_url=r.image_url,
             score=float(r.score),
         )
         for r in rows

src/chatbot/schemas.py

@@ -33,6 +33,7 @@ class SemanticSearchResult(BaseModel):
     chunk_id: str
     kb_id: str
     text: str
+    image_url: str | None = None 
     score: float
 
 class ManualTextRequest(BaseModel):